DSPeaker Anti-Mode 2.0 DualCore Digital Signal Processor
Hard though it is to believe, it is now twenty years since the first units for the digital correction of speakers in rooms began to appear. I recall that when I first wrote about the Sigtech for TAS in 1992, I called up TAS headquarters, then in Sea Cliff, and described to one of the TAS staffers (not HP, but someone who is also still around) what the Sigtech did. I shall never forget his response: “If this thing works, there will be nothing left to review.”
DSP did and does offer revolutionary power to control what one hears, though some things, speaker radiation pattern for example, are beyond its control. There are still things to review. But the Sigtech did work so remarkably well that I expected an immediate and profound impact of DSP on audio. But for various reasons, the DSP revolution moved into audio somewhat more slowly than it ought to have, considering what it had to offer. One reason as far as the Sigtech was concerned was that it was really expensive and also not all that easy to program. It was basically a professional product and the consumer had to have faith and patience to get to the startlingly good results that DSP, as embodied in the Sigtech, could offer.
But in the digital world, twenty years is a very long time indeed. And now, twenty years after my first Sigtech review, I have in hand the DSPeaker Anti-Mode 2.0 DualCore. It is reasonable in price ($1199) and simple to operate, and it offers an abundance of DSP features of great effectiveness and the highest sonic quality, including top-quality digital-to-analog conversion included in the package. The DSPeaker Anti-Mode 2.0 DualCore (just DualCore hereafter) does what it does superbly well, and it is so affordable that adding it to almost any serious audio system is financially reasonable as well as sonically beneficial in a big way. It should be irresistible to anyone who does not already have a DSP room-correction device— or even to those who do have one already, because this one is new and different. And it can vastly improve the sound of almost any audio system.
The DualCore is a digital signal-processing device. While it can be used as a straight DAC, with TosLink and USB inputs and analog outputs, it is really intended to operate on its input signals and do things to them, to make your audio results better. What it does to them is various kinds of alteration of their frequency response, i.e., equalization. It does not alter phase independently of amplitude response, although that is possible in digital signal processing (hereafter DSP). It does what is called minimum phase equalization, where the phase behavior is determined by the amplitude behavior in a particular way that mimics what most physical processes do.
The DualCore can be used in several different ways. First, it can be inserted in the tape loop of an analog preamp, with analog in and analog out, like a classical analog EQ device. In this use, the analog input is converted into high-resolution digital, processed (EQed), and then converted back into analog. Some audiophiles will be immediately worried but really they should not be: Highresolution digital A-to-D-to-A is an essentially transparent process.
Alternatively, one can give the DualCore a digital input as described above, let it convert the digits after processing into analog (which it does very well), and feed the analog output into a line-level preamp. Or one can feed the analog output directly into an amplifier: The DualCore has a volume control and can thus function as a digital preamp. The DualCore also has two analog outputs—RCA and XLR—that can drive a subwoofer system as well as the amplifier for the main speakers if desired. (More later on its use with subwoofered systems.)
Connection is straightforward and clearly explained in the manual. All you might need is a TosLink cable for your CD connection, if you do not have one. Easy to get and cheap.
The DualCore has a small but clear built-in screen that shows menus of the operations available and the results of measurements. The navigation of the menus and the choice of operations are done using a small remote control.
What the DualCore Does: Basic Bass Correction
The heart of the DualCore’s operation is a bass room-correction system. The upper limit of the frequency range corrected is user-selectable, from 80Hz up to 500Hz. The default setting is 150Hz. Why bass correction only? Because the lower frequencies are where the room has the largest effects and the largest errors consequently occur. If you have good speakers, you are already hearing something like correct response from around 500Hz on up because in that range the direct arrival from the speaker dominates what you hear. Lower down, the room has huge and audible effects that need fixing. This is almost inevitable. It is a piece of really good luck if the bass in a given room turns out perfectly. This hardly ever happens, no matter how carefully one places speakers and in how good a room. Look at the measured room responses published round and about. Almost none of them are flawless. Bass needs fixing in almost all cases. (You can see a lot of examples in my general articles on digital room correction on www.regonaudio.com.)
Of course, the domination of the direct sound above 500Hz is not total and also even the best speakers can usually use a bit of tweaking. What you can do with the DualCore above 500Hz I shall get back to in a minute.
Getting the DualCore’s bass correction going is a snap. You run the analog output of the DualCore unit into your preamp (or straight into your amplifier if you want to use the DualCore itself as a preamp). Connect the microphone, which is supplied with the DualCore, to the unit. Put the microphone where your head is when you listen (the mic has a long cord). Then you use the little hand-held remote to click through the self-explanatory menu. The unit will play a lot of bass sweeps—it takes several minutes to figure out what to do. When it is done it will show you a before and after measurement of the bass response. “After” will look a lot better! (Incidentally, as you get more advanced with the system, you will be offered the possibility of some spatial averaging—measuring at more than one position and combining the results. I am just speaking now of the basic operation.)
Then you are ready to listen. Plug any kind of input into the analog input and leave the analog output going into your amplifier—or you can put the unit input and output in the tape loop of a preamp as described above. That is probably the easiest way to get started fast, actually. The whole process is very quick and easy, no computers, no complex “learning curves.” Just follow the well-written instructions in the manual and there is almost nothing to it to get it set up.
Ah, but when you listen, there is something to it for sure. Gone are the bass booms of most systems. Precise bass with wonderful tonal and rhythmic accuracy are much in evidence. The bad effects of the room are all but perfectly eliminated.
If you took the quickest route through the menu, you are listening to correction up to 150Hz only. The next step to try is the effect of moving the top of the correction up to 500Hz and various frequencies in between 150 and 500Hz. A little experimenting here is good because exactly where you stop the correction will have some impact on how smoothly the transition to the uncorrected range occurs—and of course on exactly what kind of correction happened. It is easy to change, recalibrate, and re-listen.
There are also options available to change exactly how the correction is done. You can choose how much to have dips lifted (“Maximum” or “No Compensation” for dips are available options, with “Normal/Typical” in between); peaks are always pulled down. And if you wish, you can do corrections based on multiple measurements at different spots. There is a lot to experiment with, with further options discussed below, but it is all easy to follow using the owner’s manual and the displayed menus.
Two Examples and Some Things to Watch For
As you can see in Figure 1, the bass really does get largely flattened out by the DualCore.(unsmoothed data, as it comes from the DualCore). You can also see that the system does not attempt to fill in completely all the dips—and it is right that it does not try to do this. Often such dips are caused by cancellations of signals that are nearly in reverse polarity relative to each other, and filling them in is risky since the power increase at that frequency may have to be huge to bring the measured response up to flat. In the case of “infinite nulls,” frequencies with nearly total cancellation; the dip could not be filled in no matter what. Fortunately, deep narrow dips tend not to be seriously audible. And also if you move the speakers—and your listening position—around often enough you can get rid of them. I did not do that here because I wanted to illustrate the point about not trying to cancel them by EQ. In Figure 1, the dip at 60Hz is very deep, almost an “infinite null.”
As I said, moving the speaker can get rid of such dips in most cases and this is generally a good idea—to find a spot where deep cancellation nulls are not in evidence. Figure 1 had the speakers in positions that brought up the low bass but introduced irregularities in response from interaction with the nearby walls(Allison effect, as this is usually called).
In this case, moving the speaker to another position gave the smoother response shown in Figure 2. This is a measurement not smoothed beyond what the DualCore produces for you— just to show how well things can work if you work first at getting things in the right place.
And look at how nearly perfect the corrected response is! This is about as good a bass response, within the limitations of the speaker, as one is likely to get in listening terms. It is really smooth and the slight down-slope with rising frequency from the bottom of the speaker’s real bass extension is good in listening terms. Some people may prefer the Figure 1 position, however, for its greater extension, if reduced smoothness.
In this second position, the really low bass is still down in level as shown in Figure 2 again, but this is indeed the nature of the speaker—it rolls off starting around 50Hz anechoically. But in the speaker’s real operating range in the bass, the uncorrected response is now free of extreme dips. And after correction, the bass is essentially ideal. (This correction was done with the “Max” setting, which will be discussed momentarily.) One could pull up the bass below 45–50Hz with a “House Curve” setting, but this would run the risk of over-driving the speakers with material that had strong low-frequency content. One has to respect the limitations of the speaker itself always.
If the smooth rise as frequency decreases to 50Hz gives too much bass, you could run the correction with the “Typical” rather than the “Max” setting and get less bass. You can also turn off all compensation for dips and remove peaks only, which gives a situation with maximum headroom for “House Curve” modifications. This is all adjustable at your command!
Note that the DualCore operates primarily by reducing peaks. Room resonances (modes) usually manifest themselves as peaks. But since the system quite correctly operates by removing such resonant peaks, it is possible for the overall bass level to drop too much, especially in the “Typical” mode of operation with the filling in of dips turned off completely (and even more so with “No Compensation” for dips). Note that recordings are usually made with the expectation of some bass lift in-room: Literally flat response tends to sound lean.
In any case, the DualCore offers several ways to deal with this question of overall quantity of bass. First, as already noted, there is the choice among levels of dip filling in. A choice of “Typical,” “Maximum,” or “No Compensation” is likely to come out quite close to right. In addition, there is a feature called “Quicktone,” which is in effect various versions of the “tilt” controls made famous in analog form in Quad preamps (linear phase correction here). As also mentioned above, you can also insert a custom “House Curve,” which shelves the bass up below a userselectable frequency. There is also a more subtle way of dealing with the issue by judicious choice of the upper-limit frequency of the bass correction. This is again user-selectable.
The “calibration” is done on the summed response of two channels in the true bass. This makes sense because real bass tends to be perceived as a summed event as far as amplitude response is concerned (this is why one subwoofer works as well as it does).
You can also use the full-range measurement feature (a sweep measurement of in-room response) to check the general integration between the lower and upper frequencies. This provides a useful check on the choices discussed previously.
Note, however, that if you want to use this measurement for setting parametric EQ (see below) then you should do the measurement quite close to each speaker. This type of measurement, which gives the steady-state response, should not be flat in the higher frequencies at the listening position—smooth, yes, but not flat—but rather should slope down somewhat with rising frequency if the measurement is done at any substantial distance from the speaker. (Audio people tend to assume that any response measurement ought to come out flat, but it is the direct arrival of a speaker that ought to be flat in the higher frequencies, not the overall room sound. Measure up close for the frequencies above the bass and lower mids!)
As you see, there are a lot of ways that one can vary the detailed results of the bass correction. You should be uninhibited in experimenting with them. The whole point of all these settings and options is to let the ear be the final arbiter, as Lord Rayleigh famously said. And it is indeed useful to set up a number of options for comparison. Such comparison is easier than might be expected, because you also save various choices (“profiles”) for quick comparisons. You do not even need to recalibrate each time—you can copy a given profile into another profile setting, modify it, and then save it as a new setting. There is in addition a measurement feature that does a full-range frequency sweep so you can see how the integration of the bass correction and the frequency range on up to the top look, for example.
Clearly we are getting into a little more sophisticated set of operations now, but this never gets hard to do. This is one digital system where you will not need any outside help to get rolling—or even to get things as perfect as possible. With the easily comprehended menu, the measurement feature, and your own ears, you do not need an outside expert. Trust your ears and experiment!
More possibilities: Parametric EQ
So much for the bass. But there is yet another set of EQ settings that play an important role, potentially. The DualCore allows the user to make as many as sixteen “parametric” EQ filters, with selectable center frequency, amplitude (of lift or cut), and “Q” (width of the correction—high Q is narrow band, low Q is broader band). The device presents the bandwidth of the filter as the datum, not as Q as such. This is primarily of interest in correcting the speaker itself, though you can use it to modify the room correction as well.
As already discussed, the room affects the sound of higher frequencies less than lower frequencies because the ear/brain edits the perception of sound to emphasize direct arrival in the higher frequencies. But no speaker is absolutely perfect and most can benefit from a little touch-up—or more.
Many speakers, even high-end ones, make surprisingly large errors in response in the frequencies above the bass. And even the best can be tweaked to be better, with very few exceptions. (You might want to look at the first figure here, www.regonaudio.com/Digital%20Correction%20for%20Audio%20Part%20III.html showing what just four parametric EQ filters between 2kHz and 8kHz could do for a speaker that was already quite flat. There are very few if any speakers in the analog-only world that are as flat as this in the range that was corrected. And that is only four filters in action: The DualCore has sixteen available.)
This is a different world from purely analog speaker performance, different and better. (The corrections in the link given were actually done earlier with the Z Systems rdp-1 parametric EQ device, but similar results will be obtainable from the DualCore, parametric EQ being a mathematical and uniform process.)
The subject of how to EQ the higher frequencies of a speaker would need a long article unto itself, so I shall have to confine myself to a few summary remarks. First of all, since the ear/ brain is editing for direct arrival, one should eschew narrow-band ups and downs to get rid of the small room effects that arise from reflected sound. In essence, one should correct the speaker (in the higher frequencies) anechoically. Few of us have an anechoic chamber handy. But you can get quite accurate results by measuring near the speaker. As noted above, you can use the full-range sweep measurement of the DualCore for this.
Second, you should not correct things that are highly unstable with respect to variation of position. Spatially stable effects need corrections; unstable ones do not. And other things being equal, extremely narrow-band corrections are seldom needed with good speakers.
Again, experimenting is recommended, and again you have the opportunity to save “profiles” for quick comparisons.
The DualCore really opens up a world of possibilities. It can be up and running and correcting bass in just a few minutes. But you can also explore more of the possibilities. And even once you have made your system sound what to your ears is ideal, the Quicktone feature can be used in addition—on the fly, as it were—as a set of tone controls of a very effective sort. And none of this does any signal damage! Properly done digital EQ is penalty free, and properly done it is here.
To explore absolutely as far as one can go, it is a help to have the outboard viewing program, where one can pull the small screen pictures the unit itself shows you up on a computer screen. This will be available as a downloadable program from DSPeaker shortly, though at present it does not seem to be on the Web site.
Limitations Relative to Subwoofers
The DualCore is pre-eminently a device to correct a stereo system as a unit. While in principle one can drive a system with two main speakers and powered subwoofers using the two analog outputs, in practice one cannot set delays and high- and low-pass filters separately, as one would want to do for subwoofering in its ideal form. The DualCore as it stands will EQ a subwoofered system to be flat. But to use it as a crossover and time-delay device, it appears that one would need two of the units, as things now stand.
And even then, there would be the difficulty of changing the volumes of the two channels in a unified way. I understand from Tim Ryan, the USA distributor for DSPeaker, that plans are afoot for a double unit with a master/slave connection to gang the volume controls and thus offer full subwoofer flexibility. For the moment, be aware that while the DualCore will do an admirable job of flattening your overall subwoofered system, it will not really enable you to do the kind of high-pass and lowpass filtering with time delays that most subwoofer users would like to see in a DSP unit—not yet anyway! But stay tuned.
This review has so far been about theory and measurement and history. But of course the real reason for using a device like the DualCore is musical. High-end audio has long been obsessed with mids and lower treble, with “female vocals” being the standard test material. And of course the midrange is the heart of music. But bass counts and lower midrange counts, too. For too long, audio people have put up with the lower frequencies being wrong, and they have adjusted to this wrongness. The first time one hears the bass reproduced correctly is a revelation. Of course you might have had this revelation long ago, but in case you have not, let me describe what happens.
There is a simultaneous gain in both strength and precision of the lower midrange, upper bass, and on down into the depths. Organ music, as in, say, the Reference Recordings Rutter Requiem organ part, becomes a well-defined, pitch-accurate, and yet still powerful part of the music. Plucked string bass notes in, say, The Paul Desmond Quartet Live (Horizon/A&M) or Opus 3’s Tiden Bar Gaar begin to sound really like a string bass, with the combination of definition and impact of live bass pizzicatos.
Orchestral music gets its real foundational sound, impact, and definition combined. (Try my old stand-by, Dvorák’s New World on Delos). The change is profound in all music where bass really counts, even in something that does not seem bass-oriented, such as the string orchestra version of Bach’s Goldberg Variations, arrangement by Sitkovetsky.
The change here is so pervasive—so nearly universally does improving the bass radically improve music and increase both its realism and its musicality—that citing specific examples seem almost a diversion. Everything is better if there is bass in it at all.
And of course the possibility of fixing the small defects that almost all speakers have in the higher frequencies, that too has a pervasive and fundamental impact. In fact, the adjustment of the speaker to be flawless above the bass—from say 500Hz on up—has to my ears a more significant impact in musical terms than the exact details of bass. Getting the overall balance of the bass correct is really significant, too, but the perfection of the mids on up is the heart of audio.
Put these things all together, and a properly EQ’d system sets new standards. It is just not possible to get this level of fidelity from an uncorrected analog speaker/room combination. Very occasionally one might find a fortunate combination of room and speaker that gets close. But there is always a little bit to correct, and usually there is a whole lot to correct. And the musical impact is profound.
Moreover, there is a great increase in the perception of the space of the original venue, where the recording was made. Perhaps this is not surprising—the resonances of one’s own listening room can only obscure the acoustic signature of the recording’s venue. But the extent to which this is true may come as a surprise. It is not hard to take your room out of the listening impression in the higher frequencies. In a well-damped listening room, sitting relatively close to the speakers will do the job. But to get rid of the bass signature of your room requires something more. When this is done by removing the effects of your listening room’s irregularities, the sense of original space is greatly enhanced. (Ironic, is it not, that early high end, for which this recovery of space was a primary goal, turned its back on the most effective way of accomplishing it?)
And finally, I need to mention the marvelous advantage, especially for those toppy and often bass-shy Golden Oldies from the early days of stereo, of having effective and easily operated tone controls of the “tilt” variety, supplied in the Quicktone feature. Not a cure-all for the odder aspects of unbalanced recordings, they are, even so, a great help in moving the unbalanced into the listenable category.
History and Analog EQ: A Cautionary Tale
High-end audio decided early on to give EQ, necessarily analog as it was then, a miss. There were a few exceptions, conspicuous among them Mark Levinson’s Cello units and Dave Wilson’s EQ device included in his WAMM system. But few of the early high-end critics seemed to realize that EQ properly used was a priceless tool in audio. Part of the reason for the opposition to EQ was a lack of understanding of how it works, in particular a lot of nonsensical pontificating about how it messed up phase— when in fact minimum-phase EQ’ing to flat tends to improve phase behavior in most situations. But in addition, there seems to have been a sort of audio analog of General Jack D. Ripper of Dr. Strangelove, a kind of mad obsession with signal purity for its own sake. Never mind that the RCAs, Mercurys, Deccas, and Columbias worshipped by those same people were heavily EQed, that speaker crossovers are EQ devices by nature, and that vinyl records are EQ’d and inverse EQ’d by nature. Never mind any of that. EQ was supposed to be evil!
In view of this, it is worth noting that a surprising amount of the benefit of DSP EQ is actually obtainable by a careful use of analog EQ, even of the 1/3-octave “slider” variety that was so held in contempt by high-enders in the early days. The logic of 1/3 octave was that it is the “critical bandwidth” of the ear, the bandwidth over which the ear lumps amplitude perception together (not of course pitch but overall amplitude impression). And with a reasonable good speaker placement, one can do a surprisingly good job of correcting bass. And with a smooth speaker of correcting higher frequencies, too.
Why bring this up here? It is not because DSP EQ is not better—it is better. But, and here is the cautionary tale, if you eschew digital EQ as people formerly did analog EQ then you are going to repeat the errors of decades ago. Just as the way that for decades high-enders were mostly listening to bass that was really wrong—and higher frequencies that were not always so good either—you are going to be stuck with things that do not really work, with sound that is not really right. As to fixing bass by changing amps or cables or the like—a little experience with the DualCore will reveal the truth about that. (Imagine trying to find an amp/cable combination that will pull down a peak at 120Hz specifically!)
Back to the DualCore
The DualCore is not as automated or as elegant looking as some of its (stand-alone) competition. One can only adjust top end by the tilt controls and the parametric EQ, which takes a little patience. And one needs to tweak the bass level for ideal matching with frequencies further up. User control is good since the models of what one hears in automatic systems are never quite perfect. But, in any case, in various ways you have to do a little work at the beginning. Quite aside from the final results being good, the experimenting itself is interesting, though you have indeed to do quite a lot of it to get the best possible results. However, in the end you can get truly remarkable sound out of a system corrected and adjusted with a DualCore. One thinks of the introduction to The Three-Penny Opera: “Conceived with a splendor that only a beggar could imagine and ….so cheap that even a beggar could afford it.” Well, not quite that cheap. But still a wild bargain, and a sonic wonder.
SPECS & PRICING
Device type: Digital signal processor/digital-to-analog converter/digital preamp
Connections: Two RCA inputs,two XLR inputs, two RCA outputs, two XLR outputs, one TosLink S/PDIF digital input, one TosLink S/PDIF digital output
Digital processing: 40-bit, 2-channels
ADC: 6.144MHz oversampling dual per channel
DAC: 6.144MHz oversampling dual per channel, local clocking, buffering
Dynamic range: > 108dB
Volume control steps: 0.5dB
Input sensitivity: XLR, 1.35/2.60V RMS; RCA, 1.65/3.25V RMS
Output voltage: RCA, 1.65V RMS; XLR, 3.25V RMS
Dimensions: 236 x 55 x 146mm
Simplifiaudio (U.S. Distributor)
5415 Clairemont Mesa Blvd. #1001
San Diego, CA 92117